A 4.1 GHz Dual Core at $130 - Can it be True?

[icbluscrn]

You shouldn't allow the CPU to exceed 60C

and you should not go swimming for 1 hour after eating also.
Or eat pop rocks and soda

I had my old roomates one @78°c for like 2 months till he got watercooling.

take a look at this one (d805)prime95 still good

60° i say thats kind of cold,

 

[linux_0]

You shouldn't allow the CPU to exceed 60C

and you should not go swimming for 1 hour after eating also.
Or eat pop rocks and soda

I had my old roomates one @78°c for like 2 months till he got watercooling.

take a look at this one prime95 still good

60° i say thats kind of cold,



If it was really running @ 81C you're lucky it didn't explode.

 

[stljeffbb]

OK, here are the specs I mentioned:

Pentium D 805 2.66Mhz (of course!)
OCZ DDR2-800 2 GB RAM (dual channeled 2x 1 GB)
Asus P5WD2 (just the plain version, no E, Premium, etc.)
Zalmzn CNPS9500-LED
Sapphire X1900 Radeon w/512MB VRAM
Samsung SATA 3.0GB data / 40GB storage
Lite-On DVD
Arctic Silver 5 Thermal Compound

I also have:

Sky Hawk/Eagle Tech BAX BAX5602BK case (with side venting, but sleeve does not fit Zalman cooler)
Sunbeam 450W power supply
80mm chassis fan

..now being a noob, I'm concerned that I have my fans installed incorrectly....following the Zalman instructions, it says to put the fan blades on the side that the air flow should go (toward the back of the case), and I feel pretty good about that, however, my 80mm chassis fan came with no instructions....it certainly seems that I have good airflow coming out of the back of the 80mm chassis slot, so I think I put it in correctly....

Thanks for the replies.....

To the poster who said "you shouldn't allow the CPU to exceed 60C"....why not? In other words, if my absolute maximum temperature is 63C, why not let it run at that speed which seems to be within specs base on official Intel info?

According to the THG guide, running the stock cooler at 100% load gives a temp of 78C!!!! If this is accurate, someone who has entirely no knowledge of computers and who enjoys games (or running their computer to the max) will seemingly often go over the 70C mark, let alone 60C. Why would the Thermal Monitor 2 wait until about 80C to shut down as opposed to say 70C?

Also, my temperatures seem to be right in line with the THG specs for temps. Heck, I'm running Prime95 right now, and even with the Zalman cooler, I've briefly touched 61C at only 2.66Ghz!

Now, with this said, on my mobo, I have a 3 pin slot and a 4 pin slot for the CPU fan. When I hook up my Zalman 3 pin to the mobo 3 pin adapter, the BIOS gives me the "no CPU fan!" message. However, when I looked at the 4 pin adapter, there is a lip that oddly enough fits the 3 pin plug exactly (its almost like they wanted us to put the three pin plug on the 4 pin adapter!), so I've put it there, and the "no CPU fan" message goes away. The Zalman is definately running and seemingly running well at 2500 to 2600 RPM. However, my fan comes with the outside-the-case fan controller...I just might hook that back up to keep it running on max.

I might also consider getting a faster chassis fan, my current one running at about 2100 RPM.....

All of the above notwithstanding, is it possible that these chips can safely run above 70C? Has anyone fried their CPU (or any other part of their system) running temps like this with this chip?

To the poster who had their friend run their chip at 78C for 2 months....I'm guessing this may have reduced the life of the chip, but I certainly have no idea whether or not this is the case! 😉

Sorry for the long message,

Jeff

 

[whos_that]

Hey everyone I have my setup up an running which is the following

Gigabyte GA-8N-SLI (Rev 1.1) Bios Rev. F4
2GB PQI Turbo DDR2 667 Mem
eVGA 7600GT 256MB DDR
In an Antec Sonata II Case (has Antec 450W PS & extra 120MM fan)
Thermaltake Big Typhoon CPU Fan
Intel D 805 (the right model chip and stepping)

My problem is that I can oc my setup no problem to 3.5ghz and it seems pretty stable by running it 100% however the Easytune 5 program the mb came with which monitors my oc'ing and voltages ... etc shows that the chip voltage really fluxuates from say 1.22-1.35v ....

When I try to oc it to anything higher say 3.6 at stock voltages my system becomes unstable and I get random reboots or freezes.... so I tried a minor voltage bump and it helped but still when I max out my system (100% usage) I still get reboots or lockups but not as much....

I did try to go up to 3.8ghz with up to from 1.4 - 1.5v voltage change and my system wouldn't even go ino windows.... so I thought to just try and get 3.6 and stay with that but that isn't really working (- see above)

So I don't really dont know whats going on or what im doing wrong.... I know some of you have the same Gigabyte GA-8N-SLI board and can get 3.6 at stock voltages with no problem.... so whats the deal?

My last and only thought is that it could mabye be the powersupply? When I don't have a major load on the system the voltage monitor (easytune 5) pretty much stays at about 1.29 - 1.35v but when I tax it out it goes low say to 1.21v

thanks for you help as I don't know what to look at next or if I should buy a new PS.... if so what brand / unit / Wattage would you all recoommend?

 

[whos_that]

Hey everyone I have my setup up an running which is the following

Gigabyte GA-8N-SLI (Rev 1.1) Bios Rev. F4
2GB PQI Turbo DDR2 667 Mem
eVGA 7600GT 256MB DDR
In an Antec Sonata II Case (has Antec 450W PS & extra 120MM fan)
Thermaltake Big Typhoon CPU Fan
Intel D 805 (the right model chip and stepping)

My problem is that I can oc my setup no problem to 3.5ghz and it seems pretty stable by running it 100% however the Easytune 5 program the mb came with which monitors my oc'ing and voltages ... etc shows that the chip voltage really fluxuates from say 1.22-1.35v ....

When I try to oc it to anything higher say 3.6 at stock voltages my system becomes unstable and I get random reboots or freezes.... so I tried a minor voltage bump and it helped but still when I max out my system (100% usage) I still get reboots or lockups but not as much....

I did try to go up to 3.8ghz with up to from 1.4 - 1.5v voltage change and my system wouldn't even go ino windows.... so I thought to just try and get 3.6 and stay with that but that isn't really working (- see above)

So I don't really dont know whats going on or what im doing wrong.... I know some of you have the same Gigabyte GA-8N-SLI board and can get 3.6 at stock voltages with no problem.... so whats the deal?

My last and only thought is that it could mabye be the powersupply? When I don't have a major load on the system the voltage monitor (easytune 5) pretty much stays at about 1.29 - 1.35v but when I tax it out it goes low say to 1.21v

thanks for you help as I don't know what to look at next or if I should buy a new PS.... if so what brand / unit / Wattage would you all recoommend?

 

[linux_0]

Please start a new thread.

Thanks! :-D

 

[stljeffbb]

UPDATE......seeing that others have aimed a HBF (household box fan) at their case, I tried a variation on this theme, aiming a fairly powerful (in relative terms) smaller circular fan at the side venting in the case...AND IT WORKED! the temps for the CPU have dropped about 5C to 7C at full load...

Still, my local computer store has this "turbo" fan that sounded like a jet engine when the guy tested it.....I think I'm going to drop $15 and get it to add to the other two fans (and the outside circular fan).....this hopefully will dissapate the heat even further and allow me to do some modest overclocking (which is all that I wanted to do anyway).

Jeff

 

[Black_Knight_MC]

stljeffbb, the reason we (or at least me) usually try to stress the CPU and get it so the temp is below 60C is because of the life of the CPU. Technically, if the CPU runs at 64.1C it will last a little over the warrenty period (3 years). However, when OCing you are putting even more stress on it and to try and keep the life of the CPU long and healthy we keep it below 60C. This is also done because when playing, say games, which is the usual reason for OCing, your ram and especially video will be producing heat too, which nearly all these tests don't take into account. Since these other components are producing more heat thats more ambient heat in the case and the CPU won't be as cool. My room temperature varies from 75-85F (my room that the computer is is far from the house's temp sensor). At 75F my OCed cpu never passes 52-53C (My current setup, check my sig) and at 85F, 58-59C.

So to recap, below 60C ensures:

1. A longer CPU life (so as to sell later if only used for say a year)
2. Much more stable, as the clock rises, the same temperature becomes more unstable (2.66ghz my have a stable 63C temp but a 4.1ghz OC with 63C may be unstable)
3. Safety room, the temperature of rooms fluctuate so good to have a margin of safety. Also, temps will be even higer when other components are used in conjuciton, due to ambient case temp.

 

[stljeffbb]

Hi Black_Knight_MC....thanks for your thoughtful reply.

Based on what little I know (I am a noob), I definately agree that keeping temps down is a good idea and an important thing to do.....

I just wonder if these chips can "safely" run hotter than advertised....I mean, this is one of the premises of the THG 4.1 experiment, and based on the charts provided by THG, they intend the CPU's to run hotter than advertised, even with stock cooling!

Stock Cooler:
3.32 GHz crash 57 °C
2.66 GHz 78 °C 53 °C

Zalman Cooler:
3.80 GHz 76 °C 47 °C
3.60 GHz 74 °C 46 °C
3.32 GHz 71 °C 46 °C
2.66 GHz 64 °C 44 °C

I guess I would be concerned, knowing what I now know, that my out-of-the box CPU would run hotter than it should at peak loads.....much hotter in fact (so it seems) than the company advertised "safe" levels.

This is why I'm wondering if these temps by Intel and others are "bogus."

Nonetheless, Black_Knight_MC, your post has helped me to come to the conclusion that I must keep my temps down further, and I'll need to get some extra fan help...

Thanks,

Jeff

 

[linux_0]

stljeffbb, the reason we (or at least me) usually try to stress the CPU and get it so the temp is below 60C is because of the life of the CPU. Technically, if the CPU runs at 64.1C it will last a little over the warrenty period (3 years). However, when OCing you are putting even more stress on it and to try and keep the life of the CPU long and healthy we keep it below 60C. This is also done because when playing, say games, which is the usual reason for OCing, your ram and especially video will be producing heat too, which nearly all these tests don't take into account. Since these other components are producing more heat thats more ambient heat in the case and the CPU won't be as cool. My room temperature varies from 75-85F (my room that the computer is is far from the house's temp sensor). At 75F my OCed cpu never passes 52-53C (My current setup, check my sig) and at 85F, 58-59C.

So to recap, below 60C ensures:

1. A longer CPU life (so as to sell later if only used for say a year)
2. Much more stable, as the clock rises, the same temperature becomes more unstable (2.66ghz my have a stable 63C temp but a 4.1ghz OC with 63C may be unstable)
3. Safety room, the temperature of rooms fluctuate so good to have a margin of safety. Also, temps will be even higer when other components are used in conjuciton, due to ambient case temp.

Indeed :-D

I agree!

 

[whos_that]

Not To be A Bother But This Is In Regards To This Topic A INTEL D 805 OC'ed

 

[linux_0]

Not To be A Bother But This Is In Regards To This Topic A INTEL D 805 OC'ed

So to recap, below 60C ensures:

1. A longer CPU life (so as to sell later if only used for say a year)
2. Much more stable, as the clock rises, the same temperature becomes more unstable (2.66ghz my have a stable 63C temp but a 4.1ghz OC with 63C may be unstable)
3. Safety room, the temperature of rooms fluctuate so good to have a margin of safety. Also, temps will be even higer when other components are used in conjuciton, due to ambient case temp.

I agree with this part.

 

[stljeffbb]

Hi whos_that...

Not To be A Bother But This Is In Regards To This Topic A INTEL D 805 OC'ed

I think the cooling aspect has everything to do with OC. Why OC if you burn your chips out in 6 months? Even the article admits that cooling (or the amount of heat generated) is the "dark side" of overclocking. I mean, the entire article states how the performance of the D 805 is the same or better than chips hundreds of dollars more. However, if one can't OC and have a chip that lasts as long, then the possibility exists of having to buy a new D 805 every year or so (until the supply is exhausted....hmmm, maybe a marketing strategy?) 😉

Jeff

 

[stljeffbb]

Heh heh....well, I stopped into my local computer shop.....they claimed I had my chassis fan on backwards, that the air flow should come into the computer, however, the diagram on the Zalman instructions clearly shows that the air flow should indeed go out of the computer.....

However, upon further inspection...

I realized I put my Zalman CPU fan on backwards!

😳 😳 😳 😳 😳

Chalk it up to a noob mistake! After the fix, the temps are quite good at 2.66Ghz.....I also purchased an exhaust fan for an expansion slot for $10, and my temps have stayed hovered at 49C to 52C using Prime95....this does indeed give me some room to modestly overclock and keep those temps around 60C! 😀 😀 😀

Jeff

 

[icbluscrn]

All of the above notwithstanding, is it possible that these chips can safely run above 70C? Has anyone fried their CPU (or any other part of their system) running temps like this with this chip?

To the poster who had their friend run their chip at 78C for 2 months....I'm guessing this may have reduced the life of the chip, but I certainly have no idea whether or not this is the case! 😉

Sorry for the long message,

Jeff

well with some doubt i'd say overclocking may shorten the life of the cpu but the way i see it i have pent 1's and pent 3's that still work. so even though warranty expires after 2 years or so the cpu's last alot longer (in general) so even if they last half as long as p1's or P'3 they will have outlived there use,.... for me at least.
I dont believe if you oc to 4ghz your only get a year or less out of it. I think it more like 5 years instead of 10 o well.

So to recap, below 60C ensures:

1. A longer CPU life (so as to sell later if only used for say a year)
2. Much more stable, as the clock rises, the same temperature becomes more unstable (2.66ghz my have a stable 63C temp but a 4.1ghz OC with 63C may be unstable)
3. Safety room, the temperature of rooms fluctuate so good to have a margin of safety. Also, temps will be even higer when other components are used in conjuciton, due to ambient case temp.

I think your wrong on all counts

1. the cpu is $120 now in july its going ot drop to $90 , in one year...well i dont wont to guess what it will be after one year. Regardless I am sure life wont be cut to less then one year or two.
2. In general temp has nothing to do how stable it is, its only a biproduct of the overclock, cpu die and voltage determines more on stability because you can leave it at stock speed 2.66 and run it at 80°c on stock voltage. or you can go to 4ghz and run it at 80° why will it fail because of voltage.
3. Thats why you check for stability, alot of people on here i am guessing are oc'ing there cpu's and not checking stability I mean truely checking. you need to run/check it for hours not minutes. I like 8hours its plenty of time to stablize temps.so you know exactky where you stand
part two. The chip has Tm2 its a saftey feature its built in so there is your protection regardless

 

[stljeffbb]

Hi IcBIUsCrn....

You said:

The chip has Tm2 its a saftey feature its built in so there is your protection regardless

I guess that is one of my main questions.....why is the TM2 set so high (~80C seemingly) when Intel's specs say it should be at about 64.1C? Why not set it at 72C?

I have been testing for 2-3 hours each time before reporting figures here....maybe not quite enough time....however, even when running in the mid 70s Celsius, I have still not experienced any kind of CTD, BSOD, etc....I've even lowered the core voltage to 1.27 and change, just below the recommended value, and I've not had any difficulty.



Jeff

 

[darklife41]

With the CPU thermal control enabled in BIOS, our systems throttle back the CPU at just over 72C. I consider it safe to run at these temperatures since Intel and Asus have seen fit to set the limits there. I wouldn't run the CPU without the thermal control enabled though. Of course addtional heat will cut into the life of the CPU. However just how much remains to be seen. I'm expecting to still get 3 years out of the CPU since that's what the warranty is. Anything more is just icing on the cake as far as I'm concerned.

I think its ridiculous to run Prime95 for hours to test for stability. My temps are stable after about 3 minutes (whether its water cooled or air cooled) and I've never seen an error past 6 minutes. But to each their own. If you do get an error on Prime95 after an hour, its not likely to ever show up on any real world applications/games since nothing runs that long at this speed, let alone uses 100% of both cores to do it. Prime95's torture test has a default setting of 15mins to run each FTT. There's a reason for this. ;-) If you want to burn in your system, that's a different story.

3.6GHz on air, 4.0 GHz on water, and 4.1 GHz on water are now working links. 3.8 GHz on air is next. Will hopefully have them all done and working in another week or so. 🙂

http://www.ultramaxcc.com.au/intro.htm

 

[icbluscrn]

Hi IcBIUsCrn....

You said: The chip has Tm2 its a saftey feature its built in so there is your protection regardless

I guess that is one of my main questions.....why is the TM2 set so high (~80C seemingly) when Intel's specs say it should be at about 64.1C? Why not set it at 72C?

I have been testing for 2-3 hours each time before reporting figures here....maybe not quite enough time....however, even when running in the mid 70s Celsius, I have still not experienced any kind of CTD, BSOD, etc....I've even lowered the core voltage to 1.27 and change, just below the recommended value, and I've not had any difficulty.



Jeff

well its a long story i wish not to tell but will give you some info on later, but info on the d805 was/is hard to find in intels website but about a 2 months ago i talked to intel to see were the the d805 fit in there platform compatability guide was told the 05b series even though it was 800 class so looking at this
"For Pentium® D processors in the 775-land package, processors with different Platform Compatiblity Guides have dissipate different amounts of heat. Typically the highest speed processor in a give Platform Compatiblity Guide will dissipate the most power. (The highest speed processor will be closest ot the maximum specification.) When building systems that will feature many operating frequencies, testing should be performed using the highest frequency processor supported at the highest Platform Compatibility Guide supported, because it dissipates the most power"
I picked 67° as a good temp

Back to the 80°c Tm2 question it has to do with TDP=Thermal design power. To determine compliance to the thermal profile, a measurement of the actual processor power dissipation is required. but since toms did the watt draw it made it easier a measurement. The measured power is plotted on the Thermal Profile to determine the maximum case temperature. Keep in mind power (watts) increases linearly with
frequency and with the square of voltage. If i have not lost you let i will soon ,because i will be lost soon. The temperature at which the Tm2 signal goes active is individually calibrated during manufacturing so not all cpu's will be `80°. The power dissipation of each processor affects the set point temperature. The temperature where Tm2 goes active roughly parallels the thermal profile.
ok stay with me here i think the 80° is a low setting for this cpu because of the thermal profile plot i think its set to 80° for the thermal protection of voltage regulators. <yep even though it triggers off of the cpu temp. I did a little tooling around with Tm2 disabled.

You will also notice that on intels spec sheet they dont give a certain temp° like most of there other cpus they refere you to there own thermal profile chart.
i will also say contact intel they are a great help even with crazy questions.

Ok why is there max 64° the numbers have taking into account noise higher temps required the stock hsf to work harder ergo more noise.

 

[stljeffbb]

Cool discussion!

OK, first an update....I bought two more fans for a total of $25, and I can see a marked improvement running OCed at 3.33Ghz......

I now have a Stealth 80mm for the rear chassis, a System Blower Expansion slot fan placed below my vid card, and my Sky Hawk/Eagle Tech BAX BAX5602BK case has these cool side vents that perfectly fit an 80mm fan that I am using for intake into the case.

Using Prime95 for the last two hours, I have only once seen 60C for the CPU (not that I'm looking all of the time), and the constant CPU temp is around 57C to 58C with a mobo temp of 44C to 46C.

Heh heh, I've now got six fans running in the computer (three case, one CPU, one vid card, and one for the power supply)...still, very inexpensive to keep the thing cool and help extend the life of my chip.

IcBIUsCrn, concerning your last post, I actually did follow most of what you said.....interesting stuff! Only once have I seen my core clock/throttle memory cut back, and it was just a blip on the screen!

I'm now going to try 3.6Ghz and see what happens with my fan-o-rama-puter. This is as fast as I planned on running it anyway, and if I can keep the CPU at 65C and under with mobo of 50C, I think I will be satisified and be ready to play some games!

Jeff

 

[DexMan]

Well, After having a corrupted XP update process, I reinstalled XP pro a second time and so far so good.

System Stats:
Asus P5ND2-SLI
D805
Artic Cooling Freezer Pro 7
OCZ 667 4-4-4-12 (2GB)
Xfi Platinum
X1800 GTO
Samsung Sata 250GB
Coolermaster Aluminum Case
Fortron 600watt PowerSupply
NEC 3550a DVDRW

I still am modding the case and need to chop out the internal front grill.
Some of the things I am noting atm are that my CPU Idle Temp seems to be 29C to 31C with stock Arctic fan and paste. I'll be pulling and adding Silverstone 120 fan and Silver 5 paste tonight and I'll start pushing it to see what I can do.

 

[stljeffbb]

OK, here is what I think will be my last update.....

I've been testing my system at 3.6Ghz for about 3 hours now using Prime95.....when I first ran the tests, I received my first error statements! I have boosted the core voltage levels to 1.325 (well within specs of the chip, which is rated up to 1.4v), and this has seemingly solved the problems, as the program has run for nearly three hours without a hitch!

Temperatures have been no higher than 64C for the CPU and 49C for the mobo....I think this is as far as I want to push it, and I am quite happy with this performance!

Thanks everyone for all of your help! It's partially because of you that I was able to realize I put my Zalman CPU fan in backwards, and since I have corrected this and added a couple more fans, the performance of my rig has been most satisfactory! 😀

Jeff

 

[Black_Knight_MC]

Ok, got it installed and its been up and running for less than 45 min. Idles at 44c and max load 56-58C. Unfortunately it is only clocked at 3.2ghz. I am hoping the thermal grease is the reason and it still needs to settle in for the 200 hours. It has been about a 2-3c increase in temp for each 200mhz I OC. I might try 3.4 tonight or I might try tomorrow after the grease has had some actual time to settle in. Here is a shot doing a test:

http://www.nastynorthnyc.com/nrwebspace/BK%20stuff/805.JPG

Boy I hope I can stay below 60C @ 3.6 at least...

And Dexman, how the hell is it idling at 29-31c? Thats 84-87F, how cold is the room you are in? My ambient case temp is 36C (its about 5C higher ever since I installed my 7800GTX with AC HS, I think its not the actual temp but the fact that it vents heat right over the mobo heat sensor) and my idle was 40C stock speeds.

 

[darklife41]

http://www.ultramaxcc.com.au/intro.htm

Kind of disappointed that I haven't gotten any feedback about this. Let me know what you guys think. Any links that are broken or incorrect, how it loads on dialup (I'm spoiled with ADSL2), if its legible, etc.. The idea is to be able to set your systems up with virtually the same settings. Benchmarks should also be in the ballpark for any given setup with similar hardware. We should be able to get an idea of which MB chipsets work and which don't, or which require drastic changes, etc.. There should be variances according to video cards (which translate into more or less heat), PSUs, etc..

3.6, 3.8 and 4.0 on air are now valid links, as well as 4.0 and 4.1 on water. 🙂

http://ultramaxcc.com.au/intro.htm

 

[darklife41]

The TG article listed MBs that they used to run this OC successfully. No 945 was listed, and we have yet to see a 945 motherboard that will achieve the numbers. I suspect the 945 chipset just won't handle the bus rate. Just my guess though. For what its worth, the verdict is still out on NF4 as well. 🙂

 

[DexMan]

I have to stop underestimating the amount of time it takes to work on my computer 😛

The only thing I worked on was replacing the stock thermal paste on the Arctic 7 Pro with Arctic Silver 5, and replacing the Fan with Silverstone FM121 and fan controller. I let the system run at idle all through the night and then shut it off for the first cure this am before I left for work. When I get back home at 5:30 I'll start with the first rounds of tests. Only one I ran last night was 3dMark06 at stock 2.666cpu and vid card speeds.

3dMark06 First Test

So now I have a benchline for my overclocking.

As far as the temps in my room and the cpu idle temps, my room is not air conditioned and I think averages 70 degrees F or so. I could get a temp guage for my room and see later this afternoon if that's right .

I'm hoping the Silver 5 will drop those numbers another couple of degrees accross the board. Something else I should note is that I have fantastic airflow with my case, a Centurion 532, and I installed a 78CFM 120mm on the back, the heat gets sucked out really fast. I do admit that the fan in back needs a controller so I can slow down the rpms and quiet it down a bit.

More to follow later